Absolute Median Deviation Based a Robust Image Segmentation Model
Volume 9 Issue 1 2015
Download| Author(s): | Tahir Ullah, Lutful Mabood, Haider Ali, Noor Badshah |
|---|---|
| Abstract | The level set methods for image segmentation are usually embedded in variation framework. In the variation framework for image segmentation, a nominee level set function for capturing valuable edges and boundaries is modeled as a minimum of a well-designed functional. The design of the functional mainly bases on the image data embedded. The image data may either be the image gradients or image statistical information. The insufficient and incorrect amount of image data embedded in the functional may leads to inaccurate level set function and consequently incorrect image segmentation. In the recent work by X. F. Wang et al designed a functional (LCV) consisting mainly two terms. One of the terms was responsible for providing image global information and the other term for image local information. In continuous sense the L2 norm and in discrete sense, the statistic variance was utilized for this purpose. Although LCV model works well in some tests but it is not that much efficient and robust with respect to CPU timing and quality of detection in noisy images. To segment noisy images robustly, we propose a new model based on the concept of L1 and in discrete sense, absolute median deviation. The Experimental tests validate that the use of L1 or the statistics absolute median deviation gives best results in noisy images both in terms of quality and timing in contrast with LCV model. |
| Keywords | Level Set Methods, Absolute Median Deviation, Functional Minimization, Partial Differential Equations. |
| Year | 2015 |
| Volume | 9 |
| Issue | 1 |
| Type | Research paper, manuscript, article |
| Journal Name | Journal of Information & Communication Technology | Publisher Name | ILMA University | Jel Classification | - | DOI | - | ISSN no (E, Electronic) | 2075-7239 | ISSN no (P, Print) | 2415-0169 | Country | Pakistan | City | Karachi | Institution Type | University | Journal Type | Open Access | Manuscript Processing | Blind Peer Reviewed | Format | Paper Link | https://jict.ilmauniversity.edu.pk/journal/jict/9.1/2.pdf | Page | 13-22 |